Setting bindings using the command line

[vishakh]

Hello,

I have a test script to play with the Nillion node. I am using the docs at https://docs.nillion.com/nillion but don't see anything about setting compute or secret bindings.

Here the error I am dealing with:

Executing program..

Payments transaction hash: 6F16F44892DB5E81FF4706DB5A7BC29DA4401F9D6126DDDC38F5C91BB584ACCA
[2024-07-31T21:07:04Z ERROR nillion_client::managers::result::actions::compute::handlers] Error while rollback was executed.
Error: starting computation

Caused by:
    compute unexpected error: expected input party Party1 is not bound

Here is the program:

set -a
source /Users/vishakh/.config/nillion/nillion-devnet.env
set +a

export NODE_KEY_SEED="node_key_seed"
export USER_KEY_SEED="user_key_seed"

printf "\n\nGetting cluster information..\n\n"
nillion \
    -b "$NILLION_BOOTNODE_MULTIADDRESS" \
    --nilchain-private-key "$NILLION_NILCHAIN_PRIVATE_KEY_0" \
    --nilchain-rpc-endpoint "$NILLION_NILCHAIN_JSON_RPC" \
    --node-key-seed $NODE_KEY_SEED \
    --user-key-seed $USER_KEY_SEED \
  cluster-information \
  "$NILLION_CLUSTER_ID" \

printf "\n\n Inspecting IDs\n\n"
nillion \
    -b "$NILLION_BOOTNODE_MULTIADDRESS" \
    --nilchain-private-key "$NILLION_NILCHAIN_PRIVATE_KEY_0" \
    --nilchain-rpc-endpoint "$NILLION_NILCHAIN_JSON_RPC" \
    --node-key-seed $NODE_KEY_SEED \
    --user-key-seed $USER_KEY_SEED \
  inspect-ids

printf "\n\nStoring a secret..\n\n"
output=$(nillion \
    -b "$NILLION_BOOTNODE_MULTIADDRESS" \
    --nilchain-private-key "$NILLION_NILCHAIN_PRIVATE_KEY_0" \
    --nilchain-rpc-endpoint "$NILLION_NILCHAIN_JSON_RPC" \
    --node-key-seed $NODE_KEY_SEED \
    --user-key-seed $USER_KEY_SEED \
    store-values \
    --secret-integer foo=3 \
    --cluster-id "$NILLION_CLUSTER_ID" \
    --ttl-days 365)
echo "$output"
STORE_ID=$(echo "$output" | grep "Store ID:" | awk '{print $NF}')
echo "Extracted Store ID: $STORE_ID"

printf "\n\nRetrieving secret..\n\n"
nillion \
    -b "$NILLION_BOOTNODE_MULTIADDRESS" \
    --nilchain-private-key "$NILLION_NILCHAIN_PRIVATE_KEY_0" \
    --nilchain-rpc-endpoint "$NILLION_NILCHAIN_JSON_RPC" \
    --node-key-seed $NODE_KEY_SEED \
    --user-key-seed $USER_KEY_SEED \
    retrieve-value \
    --cluster-id "$NILLION_CLUSTER_ID" \
    --secret-id foo \
    --store-id "$STORE_ID"

printf "\n\nWriting a program to disk\n\n"
rm /tmp/double.py
echo "
from nada_dsl import *

def nada_main():
    party1 = Party(name=\"Party1\")
    a = SecretInteger(Input(name=\"foo\", party=party1))

    result = a + a

    return [Output(result, \"my_output\", party1)]
" | tee /tmp/double.py

printf "\n\nCompiling Nada program..\n\n"
pynadac --target-dir /tmp --generate-mir-json /tmp/double.py

printf "\n\nStoring a program..\n\n"
output=$(nillion \
    -b "$NILLION_BOOTNODE_MULTIADDRESS" \
    --nilchain-private-key "$NILLION_NILCHAIN_PRIVATE_KEY_0" \
    --nilchain-rpc-endpoint "$NILLION_NILCHAIN_JSON_RPC" \
    --node-key-seed $NODE_KEY_SEED \
    --user-key-seed $USER_KEY_SEED \
    store-program \
    --cluster-id "$NILLION_CLUSTER_ID" \
    /tmp/double.nada.bin double)
echo "$output"
PROGRAM_ID=$(echo "$output" | grep "Program ID:" | awk '{print $NF}')
echo "Extracted Program ID: $PROGRAM_ID"

printf "\n\nExecuting program..\n\n"
nillion \
    -b "$NILLION_BOOTNODE_MULTIADDRESS" \
    --nilchain-private-key "$NILLION_NILCHAIN_PRIVATE_KEY_0" \
    --nilchain-rpc-endpoint "$NILLION_NILCHAIN_JSON_RPC" \
    --node-key-seed $NODE_KEY_SEED \
    --user-key-seed $USER_KEY_SEED \
  compute \
    --cluster-id "$NILLION_CLUSTER_ID" \
    --store-id "$STORE_ID" \
    --result-node-name "my_output" \
  "$PROGRAM_ID"

Is there a way to set bindings using the command line?

Thanks!

[wwwehr]

My colleague @mfontanini says:

pass in --bind-dealer Party1=<your-party-id>

this should do it!

when building harnesses or workflow around compute circuits, I highly recommend that you take a few minutes to set up the nada workbench. it's super convenient and fast to rinse+repeat.

[wwwehr]

also consider that python is a much more robust language than shell with all the SDLC tooling you could ever want - while still suitable for scripting; as opposed to shell that is less forgiving.

[vishakh]

Thanks, @wwwehr.

As discussed, all our core work is being done using the Python and JS SDKs.

We are looking at the command line client for our DevOps options.